From: Green, Simon (Simon.Green@EU.ALTRIA.COM)
Date: Thu Nov 13 2003 - 08:00:28 EST
> From: -ray
<SNIP>
> The first config i did looks like this and seems to work fine:
>
> 1st adapter A2->8, A1->9, 2nd adapter A2->1, A1->16.
>
> But the "Understanding SSA Subsystems" book seems to suggest
> this cabling
> for a "Simple loop with two adapters":
>
> 1st adapter A2->8, A1->9, 2nd adapter B2->1, B1>16.
>
> Any advice as to which is more correct?
I believe that both are equally correct. Personally, I would do
it your way: I like to keep my A and B loops discrete.
There MIGHT be something to be said for splitting the drawer: set
the 1/16 and 8/9 jumpers to forced in-line and have two separate
loops, with mirroring between them. This can protect you from
some hardware failures. In that case, you could use the B loop
as well: have the front of the drawer in the A loop and the back
in the B loop. However, the sorts of problems that make this
worthwhile are very rare, and it would make future expansion more
difficult. (Some disk failures can take out an entire loop: I've
seen this once in six years working with a large and complex SSA
configuration. In theory, one of the SSA Planars in the drawer
could fail; I've never seen this.) And of course, more cables
means more expense, more complexity and more components to fail,
(although cable failures are pretty rare, too.)
If you're using a 160Mb adapter and disks I don't think you'll
see any performance benefit using both the A and B adapters with
only a single drawer in the loop. You might get some benefit if
you're using an older type of adapter/disks, or of course with
two or more drawers. It might give some improvement if you've
got a _really_ I/O intensive application.
Simon Green
Altria ITSC Europe Ltd
AIX-L Archive at https://new-lists.princeton.edu/listserv/aix-l.html
AIX FAQ at http://www.faqs.org/faqs/aix-faq/
N.B. Unsolicited email from vendors will not be appreciated.
Please post all follow-ups to the list.
This archive was generated by hypermail 2.1.7 : Wed Apr 09 2008 - 22:17:21 EDT